AI versus Marshy #55 - 3 uncomfortable truths about AI right now, Aussies killing it, and tools down


Hello and welcome to another edition of AI versus Marshy!

Welcome to our new readers - nice to see you 👋

This is the newsletter that keeps you abreast of the AI hype machine, and grounds it in - yeah but what’s the reality now?

This week looks at:

  • 3 uncomfortable truths about AI right now
  • 3 Aussie companies actually doing things with AI
  • The (last) update on TheLeadMagnet.biz (in current format)

Lots to run with today, so let’s make like a sprinter and leg on with it 🏃🏽‍♂️💨

-Marshy

3 uncomfortable truths about AI right now

I believe AI will eventually match the hype.

Yet there’s some cold hard truths about present state that need to be faced.

These are real issues today, and without these issues being worked through, there won’t be a healthy collective of humans to enjoy the benefits we’re being pitched every day.

1. We are probably in a massive bubble.

Crashes are cyclical. Tech is heading for one again. This is an amusing interview about the bubble with Ed Zitron.

The interview is entertaining - Ed pokes holes in how companies like Meta, Google, and Salesforce are pitching AI on their earnings calls versus what the reality is.

For even more railing - you can see Ed’s latest post with over 60 examples of crappy behaviour and broken promises - most of which are from the last couple of years.

For me - the poking with glee is schtick that gets tired after a while.

Tech has a reputation for bouncing back famously after a burst, and I’m pretty sure this is going to happen again.

2. Putting chat into AI is horrible on its own

Benedict Evans does a deep-dive on what’s happening in AI right now by doing one of my favourite things - looking at historical numbers.

ecommerce started around 2003, but took time for adoption to spread - pesky things like supply lines, logistics, bandwidth, and consumer behaviour all had to shift before it enjoyed prime time.

The same happened with the iPhone and its likely this is what is happening with ChatGPT now.

The initial surge of interest hasn’t equated to returning users just yet.

Modern smartphones are almost indispensable to most people today with their utility - but most of us have forgotten just how shitty the original iPhone was.

That’s what ChatGPT (and other LLMs) are now.

It’s going to take time for this technology to be truly indispensable.

3. The real problem isn’t an existential threat - its corporate greed.

The existential threat argument goes along the lines of:

  1. We’re currently the smartest
  2. Computers are useful, but can’t reason naturally like we do
  3. When/if they can reason - they will then be able to grow exponentially in intelligence that far surpasses ours
  4. We’ll die.

There’s a classic thought experiment referenced a lot involving paperclips - AI will keep finding better ways to create paperclips - including wiping out humans if they get in the way of creating them (it’s a published paper and I’m doing it injustice).

The real threats to humanity are much more grounded.

Timnit has a spicy take on the real safety issues in this LinkedIn post - with more information on the issues with data gathering and what’s involved in feeding AI in the Data Worker’s Inquiry.

I’ve already mentioned Google’s abandonment of its own climate agreement in pursuit of fuelling AI, and there’s more information on this to be found in this article: Generative AI is a climate disaster.

I’ve tried to dig into this issue and information is scant.

The tech companies aren’t going into the detail on water or power usage since it’s arrived:

It’s safe to say this is currently a bigger issue than killer paperclip robots.

3 companies out of Australia actually doing things with AI

I straddle the startup world a lot - with most of my coaching being focused on business owners at this stage.

Because of this visibility, here’s 3 Aussie companies you might not have heard of yet - and if you do in the future remember this moment!

  1. Andromeda

An AI and robotics startup focused on building compassionate human-robot interactions. You can follow their first robot (Abi) on Insta too!

Source video: Andromeda website

  1. Phonely

Want AI to answer the phone for you? Well there’s an app for that! Will is one of the co-founders I met at StartSpace and has now moved into the latest YC Combinator round. I played with an early version of the product and according to Will’s updates the tech that powers this space keeps getting better.

  1. InTruth

Most of our decision-making is emotional, but as humans we’re terrible at recognising this. This tech aims to change this, and is now moving into commercialisation stage with their first partner signed in Biostrap.

The last TheLeadMagnet.biz update (for now)

So I’m up to my 5th sprint and validated things as far as I wanted to.

The final numbers:

  • 655 visitors
  • 3 “sounds good”
  • No sign-ups
  • $250 in ad spend

That’s enough validation - it seems like I had no trouble getting people to the page (good ads) and a hard time getting people to convert (bad offer).

So I’m taking a different tack with this (and reworking the offer) to make this pop a different way.

Watch this space and thanks for humouring me :)

I don’t talk about them very much on the newsletter, but I am a Western Bulldogs (Australian Rules football team, AFL) fan and am starting to get excited about the end of the season.

I’ll be taking a break from day-to-day responsibilities and driving over to Geelong to watch them play an Away match in the cold on Saturday night.

I hope your finish to the week also features something you’re doing just for you!

Take care, reply with any questions, and remember we’ve got this!

-Marshy

AI versus Marshy

I call out big tech company bullsh*t, avoid hype, and show scaling companies how to grow with AI.

Read more from AI versus Marshy

Hey readers, We’re now two weeks away from ChatGPT being 3 years old. It’s safe to say awarness is close to “your uncle telling you to buy bitcoin” stage now. I’ve done my best to cover it. Partly out of morbid curiosity, partly out of bemusement, and partly out of the belief that this technology is (eventually) going to fundamentally change everything. Yet clashing with that belief is a recurring question: “Are we fucked?” Why does it keep popping up? Well - have you seen the movie The Big...

marshy smiling

Hello stranger! Welcome to another edition of AI versus Marshy. Something unfortunate happened since last edition - I started writing a “let’s go deep on a topic” edition to change things up. Then it got too large, and the tool I was covering wasn’t as easy to use as originally hoped, and things got stuck. So then I abandoned ship, got really busy with client work (good), then went ahead and started writing another edition covering Claude Skills and some other things and then go stuck again!...

a man tying a knot against a fence

Ahoy and welcome to another edition of AI versus Marshy! It’s the newsletter that charts AI through the lens of impact, growth, and technology. It’s learn-in-public style, readers say they love the deep-dives, opinions, and dissembling of information that takes place here because it’s genuine, thought-out, and not full of crap. I was thinking about what to write about this week, and thought I’d put to paper some predictions based on what I am seeing as a freelance consultant. Like all...